An undetermined number of Claude conversations and Artifacts were discovered publicly searchable on Google over the weekend, after Reddit users found that search operators like site:claude.ai/share surfaced a long list of shared conversations2.
Some of the exposed chats reportedly contained health records, private company documents, and the names and phone numbers of children. The chats were also found indexed on Bing3.
The issue appears to have originated from Claude's "share chat" feature, which allows users to create links that enable anyone with the assigned URL to view a conversation or project. Claude's interface warns that "Anyone with the link can view," language that TechCrunch noted implies the feature is intended for sharing with friends, colleagues, and small groups — not the open internet. TechCrunch drew a comparison to Google Docs, which offers a similar sharing feature but whose documents do not end up publicly accessible on Google.
Anthropic appeared to attribute the exposure to user behavior. The company told TechCrunch that share links only appear in search results when they have been posted somewhere search engines can see, such as a forum or social media post.
Wired characterized the incident as illustrating the difficulty of preventing web crawlers from making ostensibly private AI chatbot conversations public. ZDNet published guidance on how affected users should handle the situation1.
ANALYSIS The distinction Anthropic drew — that links were indexed only after users posted them in crawlable locations — does not address the core design question of whether shared links should have been crawlable by search engines at all. The Google Docs comparison raised by TechCrunch is instructive: that product uses technical measures to prevent indexing of shared-but-unlisted documents, a step that was evidently not in place for Claude's share feature. The presence of sensitive personal data — including children's information and health records — in the indexed results elevates the incident beyond a technical misconfiguration into potential regulatory and reputational exposure for Anthropic.